基于类神经网络模型的电路自主修复方法

Self-repairing Method of the Electronic Circuit Based on Neural Network Model

为进一步提高电子系统在恶劣环境下的生存能力,对利用演化硬件实现电路自主修复的方法进行了研究。首先,根据FPGA芯片与多层前馈神经网络的相似性,建立了一个可用于数字电路演化的门级电路模型,设计了专用的二进制列向量编码方法;然后,给出了实现电路自主修复的工作流程,探讨了进行电路故障诊断和修复的途径,提出了快速重构与演化相结合的电路修复方法。Matlab仿真试验表明：基于类神经网络模型的遗传算法较适合进行演化修复操作;修复效果分析表明：快速重构与演化修复相结合的修复方法比单纯依靠演化修复更为便捷。

For the purpose of improving the viability of electronic system in atrocious environment,this paper studies the circuit self-repairing method via EHW（Evolvable Hardware）.Firstly,a gate-level circuit model which can be used in digital EHW is built according to the similarities between FPGA chips and the multi-layered feed forward neural network,and a binary column vector encoding strategy is designed.Afterwards,the work flow of the self-repairing system is presented,the approaches of fault element diagnosis and circuit repairing are discussed,and a circuit repairing method through fast reconstruction and evolution is proposed.Matlab experiment results indicate that the genetic algorithm based on the neural network model is suitable to repairing through circuit evolution;related analysis proves that the method through fast reconstruction and evolution is more convenient and smarter than the method through evolution only.